AI in dating: what the apps have built, what singles think, and what could go wrong

In February 2026, Grindr began testing a premium subscription built around its own artificial intelligence. During the pilot, prices were randomised across eligible users and, according to Global Dating Insights, reached as high as $499 a month for some. It was an extreme number, but a telling one. The dating industry has decided that AI is worth charging for.

The rest of the sector has moved more quietly. Tinder reads camera rolls (with permission). Hinge coaches your opening line. Bumble is building a matchmaker you talk to. Match Group now tells shareholders it wants to become “an AI-native company”.

What follows is a sober look at AI in dating as of October 2026: the features that verifiably shipped, what the survey data says singles want, and the risks that come with machines on both sides of the conversation.

AI dating features that actually launched

Announcements are cheap; we have listed only features with a confirmed launch or test.

App	Feature	What it does	Status (per company or press)
Tinder	Photo Selector	Picks 10 profile photos from your camera roll	Launched in the US, July 2024
Tinder	Chemistry	Learns from your camera roll, with permission, to surface a few relevant profiles daily	Live in New Zealand and Australia, late 2025
Hinge	Prompt Feedback	Rates your prompt answers and suggests improvements	Launched January 2025
Hinge	Convo Starters	Offers three tips for a first message	US, December 2025; UK, February 2026
Bumble	Deception Detector	Flags spam, scam and fake profiles	Announced February 2024
Bumble	Bee	Conversational AI matchmaker	Announced March 2026, in pilot
Grindr	A-List, Discover, Profile Insights (EDGE tier)	Chat recaps, daily picks, compatibility signals	Testing from February 2026

Tinder’s AI: from photo picker to camera-roll matchmaker

Tinder’s first mainstream AI tool was its Photo Selector, launched to US users on 17 July 2024. It uses facial detection from a selfie to find you in your camera roll, then proposes 10 photos it predicts will perform well, skipping group shots and filtering out anything that breaks the rules.

Chemistry goes further. In its third-quarter 2025 shareholder letter, Match Group described it as an AI-powered feature that “with permission, learns from their camera roll to better understand their interests and personality”, using deep learning to combat “swipe fatigue” by surfacing a few highly relevant profiles each day. It was live in New Zealand and Australia, with expansion planned.

By August 2026, Match’s second-quarter letter said Tinder was testing “AI-powered profile building” as part of a reimagined profile. Read our Tinder review for how the app works today.

Hinge’s AI coaching: nudges, not ghostwriting

Hinge has been careful to frame its tools as coaching. Prompt Feedback, launched on 15 January 2025, rates prompt answers as “Great Answer”, “Try a Small Change” or “Go a Little Deeper”. It does not write the answer for you. Hinge said 63 per cent of surveyed users struggled to know what to put on their profiles.

Convo Starters followed on 8 December 2025 in the US and reached the UK in February 2026. After you like someone, it offers three tips tied to a specific photo or prompt; you still write the message yourself, and the feature can be switched off. Hinge said 35 per cent of daters in early testing felt more confident reaching out. In its Q3 2025 letter, Match said the feature drove roughly 10 per cent more likes with comments during testing.

The same letter said the long-running “Are You Sure?” prompt, which asks people to reconsider a potentially offensive message, had been rebuilt with large language models. Our Hinge review covers the rest of the app.

Bumble’s AI bet: from scam filter to matchmaker

Bumble’s Deception Detector, announced in February 2024, was defensive AI. In testing, Bumble said it automatically blocked 95 per cent of accounts identified as spam or scams, and user reports of spam, scams and fake accounts fell 45 per cent in the first two months. Bumble’s own research found 46 per cent of women felt anxious about whether matches were authentic.

The more ambitious project is Bee, unveiled alongside Bumble’s fourth-quarter results on 12 March 2026. TechCrunch reported that Bee is an assistant you type or speak to, which learns your values, goals and communication style through private conversation, then powers a feature called “Dates” that pairs people and explains why. At announcement it was in an internal pilot. More in our Bumble review.

Grindr’s AI: the wingman becomes a premium tier

Grindr’s “AI wingman” had been promised for nearly two years when, in August 2025, Fast Company reported that its A-List feature, which summarises past chats and sorts contacts, had been built with Amazon Web Services, Anthropic’s Claude Sonnet and Meta’s Llama models, and was live for a quarter of Unlimited subscribers.

Those tools now sit inside EDGE, the tier piloted in Australia and New Zealand from February 2026. In its Q2 2026 shareholder letter, Grindr called EDGE its “AI-native premium tier” and said it expects it to be “one of the primary drivers of revenue growth in 2027 and beyond”. See our Grindr review.

What Match Group tells investors about AI

Shareholder letters are where the rhetoric is most candid. In May 2026, Match said: “We see AI as a core enabler of improving user outcomes, enhancing product experiences, increasing relevance,” and announced a company-wide AI programme “with the goal of becoming an AI-native company”. It also said it would reduce headcount growth for the rest of the year with AI in mind.

In August, it added that “AI has accelerated execution across product development lifecycles” and that Hinge was building “its first reinforcement-learning model to better understand when a user may need help”. Grindr went further on productivity, estimating that engineering output rose about 2.5 times between July 2025 and April 2026 with minimal growth in its technical team.

Investors, in other words, are being sold AI as efficiency as much as romance.

How many singles use AI for dating? The survey evidence

Most of the data here is commissioned by companies with an interest in the answer. We label each one.

Match’s Singles in America 2025 (company-funded, run with the Kinsey Institute, fielded by Dynata, 5,001 US singles aged 18 to 98). Published in June 2025, it found 26 per cent of singles were using AI to enhance their dating lives, a 333 per cent increase on the previous year. Nearly half of Gen Z singles had used it; 44 per cent wanted AI to help filter matches and 40 per cent wanted help with their profile.

Match Group’s 2026 AI research (company research, about 1,000 US singles aged 18 to 39, fielded 13 April to 8 May 2026). Its June 2026 findings were more ambivalent: 64 per cent could see AI helping their dating life, but 47 per cent viewed AI in romantic contexts negatively. The jobs singles would hand over are practical: keeping a conversation going (27 per cent), building a stronger profile (27 per cent), starting a conversation (26 per cent) and planning a date (24 per cent). For deeper personal advice, 60 per cent turn to friends and 20 per cent to AI.

Norton’s 2025 Cyber Safety Insights Report on Online Dating (commissioned by Gen, Norton’s parent, fielded by Dynata, 1,001 US adults, 5 to 16 December 2024). The February 2025 release found 63 per cent would use AI to develop their dating profiles and 64 per cent were interested in AI-written pick-up lines. It also found 60 per cent of current online daters believed they had already chatted with someone whose messages were AI-written.

Put together, the pattern is consistent. Singles are happy to let AI polish the shop window. They are much less comfortable when it starts doing the talking.

Deepfakes, catfishing and AI-written messages: the risks

The same technology that tidies a profile can fabricate one. In its July 2024 deepfake research (2,000 UK residents aged 16 and over, June 2024), Ofcom found 43 per cent had seen at least one deepfake online in the previous six months.

In dating, the threat is fraud. City of London Police said in May 2026 that AI-generated images and messages were being used to support fake identities in romance scams, as reported losses passed £102 million. The Norton survey found only 46 per cent of people could spot a fake image when tested, and 27 per cent of dating app users said they had seen their own photos on someone else’s profile. Our romance scam statistics report sets out the scale.

The apps’ answer is, increasingly, more AI. Match said Tinder’s Face Check, which uses a short video selfie to confirm a user is real and matches their photos, produced a 60 per cent reduction in views of profiles later identified as bad actors and a 40 per cent fall in reports of bad-actor activity. By May 2026, Match said Face Check was fully rolled out in markets including the US, UK, Australia and Canada, and had reduced interactions with bad actors by 20 to 30 per cent “with minimal impact on revenue”.

There is a privacy trade-off. Chemistry is opt-in, but it asks for exactly the material, a camera roll, that most people guard most closely.

The rise of AI companions

At the far end of the spectrum, some people are not using AI to find a partner but as one. Pew Research Center’s February 2026 survey of 5,119 US adults found 10 per cent use AI chatbots for emotional support or advice and 4 per cent for companionship.

Among younger singles the figure is higher. Match Group’s 2026 research found 12 per cent of 18 to 24-year-olds had tried a companion app in the previous three months, mostly for entertainment or role-play. Yet two in five singles said they would refuse to date someone who used one, rising to 51 per cent of women aged 18 to 24.

Regulators are watching. On 11 September 2025, the FTC opened an inquiry into AI chatbots acting as companions, sending orders to seven companies including Character Technologies, Meta, OpenAI, Snap and xAI, focused on harms to children and teenagers.

Our AI companions hub reviews the leading apps for adults, with separate guides to AI girlfriend apps and AI boyfriend apps.

What AI in dating means for you

On the evidence so far, the most useful AI in dating is the least glamorous: tools that flag scams, verify faces and nudge you towards a better first line. Hinge’s insistence that you write your own message is the right instinct, and the survey data suggests singles agree.

The bigger bets, a matchmaker you talk to or an algorithm that reads your photo library, are still in pilots. Whether they reduce fatigue or simply add a subscription tier remains to be seen. For a broader view of where apps are heading, see our pieces on dating app fatigue and our ranking of the best dating apps.

Frequently asked questions

Which dating apps use AI?

Tinder (Photo Selector, Chemistry, Face Check), Hinge (Prompt Feedback, Convo Starters), Bumble (Deception Detector, with its Bee assistant in pilot) and Grindr (A-List and other tools in its EDGE tier) have all launched or tested AI features.

How many people use AI for online dating?

Match’s 2025 Singles in America study, which Match funds, found 26 per cent of US singles were using AI to enhance their dating lives. A 2026 Match Group survey of under-40s found 64 per cent could see AI helping.

Can dating apps detect AI-generated fake profiles?

They are trying. Bumble said its Deception Detector blocked 95 per cent of identified spam and scam accounts in testing, and Match says Face Check video-selfie verification cut views of bad-actor profiles by 60 per cent on Tinder.

Is it wrong to use AI to write dating messages?

Not inherently, but many people dislike it: 47 per cent of singles in Match Group’s 2026 research viewed AI in romantic contexts negatively. Using AI for ideas, then writing in your own voice, is the approach Hinge has built its tools around.

Are AI companions replacing dating?

Not on current data. Pew found 4 per cent of US adults use chatbots for companionship, and two in five singles in Match Group’s research said they would not date someone who used a companion app.

Sources

• Global Dating Insights, “Grindr Tests Premium AI EDGE Subscription Tier” (11 February 2026): https://www.globaldatinginsights.com/featured/grindr-tests-premium-ai-edge-subscription-tier/
• TechCrunch, Tinder Photo Selector launch (17 July 2024): https://techcrunch.com/2024/07/17/tinder-ai-photo-selection-feature-launches
• Match Group, Q3 2025 shareholder letter (4 November 2025): https://www.sec.gov/Archives/edgar/data/891103/000089110325000179/mtch8-k20251104ex992.htm
• Match Group, Q1 2026 shareholder letter (5 May 2026): https://www.sec.gov/Archives/edgar/data/0000891103/000089110326000072/mtch8-k20260505ex992.htm
• Match Group, Q2 2026 shareholder letter (4 August 2026): https://www.sec.gov/Archives/edgar/data/0000891103/000089110326000128/mtch8-k20260804ex992.htm
• Hinge, Prompt Feedback (15 January 2025): https://hinge.co/press/prompt-feedback
• Hinge, Convo Starters (8 December 2025, updated February 2026): https://hinge.co/newsroom/convo-starters
• TechCrunch, Bumble Deception Detector (5 February 2024): https://techcrunch.com/2024/02/05/bumbles-new-ai-tool-identifies-and-blocks-scam-accounts-and-fake-profiles
• TechCrunch, Bumble introduces AI assistant Bee (12 March 2026): https://techcrunch.com/2026/03/12/bumble-introduces-an-ai-dating-assistant-bee/
• Fast Company, Grindr’s AI-native plans (28 August 2025): https://www.fastcompany.com/91393167/grindr-gai-gay-i-ai-native
• Grindr, Q2 2026 shareholder letter (6 August 2026): https://www.sec.gov/Archives/edgar/data/0001820144/000182014426000023/shl_q2x2026xfinal.htm
• Match and the Kinsey Institute, Singles in America 2025 (10 June 2025): https://match.mediaroom.com/2025-06-10-Match-and-The-Kinsey-Institute-Unveil-14th-Annual-Singles-in-America-Study
• Match Group, “AI is everywhere, romance is still off limits” (18 June 2026): https://mtch.com/news/ai-is-everywhere-romance-is-still-off-limits/
• Gen/Norton, “Romance Reimagined: How AI is Playing Cupid and Catfish” (4 February 2025): https://newsroom.gendigital.com/2025-02-04-Romance-Reimagined-How-AI-is-Playing-Cupid-and-Catfish
• Ofcom, “Deepfakes: demean, defraud, disinform” (23 July 2024): https://ofcom.org.uk/online-safety/illegal-and-harmful-content/deepfakes-demean-defraud-disinform
• City of London Police, romance fraud figures (5 May 2026): https://www.cityoflondon.police.uk/news/city-of-london/news/2026/may/romance-fraud-costs-uk-victims-102-million-in-a-year-as-reports-surge-by-nearly-a-third
• Pew Research Center, “Americans and AI 2026” (17 June 2026): https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/
• FTC, inquiry into AI chatbots acting as companions (11 September 2025): https://www.ftc.gov/news-events/news/press-releases/2025/09/ftc-launches-inquiry-ai-chatbots-acting-companions